Celestia Forums

From experimenting in VB.net, I have discovered the following. Note that this may be incorrect, if so please feel free to correct me.

The format given in the helpfile for the stars.dat file is incorrect.

Two catalog numbers are stored, the format of the record is:

4-byte int: Hipparcos catalog number
4-byte int: Henry Draper catalog number (this was omitted in the description)
4-byte float: Right Ascension
4-byte float: Declination
4-byte float: Parallax
2-byte int: Apparent magnitude multiplied by 256
2-byte int: Stellar class
1-byte: Parallax error

-----

Stellar class is apparently divided up into sections of four bits:

bits 0-3: Luminosity code:

1 = Ia
2 = Ib
3 = II
4 = III
5 = IV
6 = V
7 = VI

bits 4-7: Spectral subdivision (the "2" in G2V)

0 = 0
1 = 1
...
9 = 9

bits 8-11:
0 = O
1 = B
2 = A
3 = F
4 = G
5 = K
6 = M
(I'm not good on spectra below M, the following are really just guesses)
7 = R
8 = S
9 = N

-----

What I don't know is how Wolf-Rayet stars, white dwarfs and neutron stars would be specified under this scheme.

Oh yes, and the distance for Achird (Eta Cassiopeiae) is wildly out - over 1400 light years???

chaos syndrome wrote:Oh yes, and the distance for Achird (Eta Cassiopeiae) is wildly out - over 1400 light years???

You must have an old version. The distance in the current stars.dat has been corrected.
See http://members.fortunecity.com/guilpain/index_uk.htm for more information on the format of stars.dat.

Grant

Thanks for the pointer to the webpage - although the sidebar does not link to the English version of the stars.dat help-page, it does exist!!!

About this new version of stars.dat, I'm using the file bundled with the Windows executable of Celestia 1.3.0

Anywhere I can get the new stars.dat, short of redownloading the entire program?

chaos syndrome wrote:Anywhere I can get the new stars.dat, short of redownloading the entire program?

Download it from the Celestia CVS tree at http://cvs.sourceforge.net/viewcvs.py/celestia/celestia/data/stars.dat

Grant

Thanks. Achird now is in the right place

After checking out that page, I'm still a bit confused:

If the object type is specified as white dwarf, neutron star or black hole, how do the other three data fields affect the star?

As far as I know, these objects do not fit into the OBAFGKMRSN classification scheme.

chaos syndrome wrote:If the object type is specified as white dwarf, neutron star or black hole, how do the other three data fields affect the star?

The don't do anything at present - though Chris and I have discussed ways of using the other three data fields to provide spectral details of white dwarfs.

Grant

Ive been messing around with creating stars.dat files recently.

I found this code useful.
It is from celengine/stellarclass.h, it defines the 2byte stellar class.

t is the star type (normal, white dwarf, blackhole)
sc spectral class
ssub spectral sub class
lum luminosity.

hope this helps.

Is there any possibility to add color information to the stars.dat in future versions? So the stars would look right?

There's already a look-up for temperature according to spectral class - that would allow the derivation of a colour that matches the blackbody spectrum for the given temperature, without having to increase the size of stars.dat.

Grant

True, the difference would hardly be notable...
I just would like to see the colors more accurate than they're now. Those pink stars drive me mad!

I've already floated the idea of colouring the star texture according to blackbody temperature on the Celestia developer's list (I even included a nifty chart of colours for different temperatures) but no-one took me up on it - Chris obviously has a stack of tasks he's got to prioritize, and since I can't program in C it's not a task I can take on myself.

One problem is that there is no single "realistic" colour for the stars - I'd say there are at least three:
1) The approximate blackbody colour, which would be visible from some reasonable fraction of a lightyear out to several lightyears, depending on the star's luminosity.
2) A "low colour detail" shade, which would be visible when the star is so far away it's too faint to properly excite all the eye's colour receptors, but is still not so dim as to appear white. The noticeable difference in this range is that cool stars appear very red, like Antares.
3) The "fully saturated colour receptors" shade, which would be visible at normal planetary distances, and which would be white for pretty much all stars except brown dwarfs. (At high intensity, almost any blackbody visible spectrum looks white - the filament in your lightbulb is at 3000K, the same temperature as an M-type star.)

But if we introduced such realistic colouring, I imagine we'd need to added several new entries to the FAQs, since many people would howl because "red" stars weren't red any more.

Grant

granthutchison wrote:There's already a look-up for temperature according to spectral class - that would allow the derivation of a colour that matches the blackbody spectrum for the given temperature, without having to increase the size of stars.dat.

Grant

Ok, thanks for the help. I now have a program which will search out stars according to spectral type.

According to http://members.fortunecity.com/guilpain/Stars_uk.htm, type code 10 represents WC and 11 represents WN.

These spectra are of Wolf-Rayet stars. Now I go and find examples of these stars and look them up in Celestia.

HIP 37791 is an example of a WC type Wolf-Rayet star.
HIP 5100 is an example of a WN type Wolf-Rayet star.

Now if you Go To one of these stars, please observe the temperature of the star.

HIP 37791 is listed as 3370K, while HIP 5100 is listed as 3580K.

Now visit this site: http://cfa-www.harvard.edu/~pberlind/atlas/htmls/wrstars.html.

The typical temperatures for Wolf-Rayet stars are in the 25000-50000K range.

Plus a feature request to display the spectral types of these stars as WC and WN rather than the generic W shown in Celestia when you have the text display on.

Or am I fundamentally misunderstanding the WC and WN spectral types?

Thanks,

Chaos Syndrome[/url]

chaos syndrome wrote:HIP 37791 is listed as 3370K, while HIP 5100 is listed as 3580K.

I think you're really going to have to upgrade - in 1.3.1pre11 these stars are assigned very reasonable temperatures of 47000K and 50000K. (The quoted radii are pretty suspect, but that's a symptom of the likely error in distance.)

Grant